Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.

Big Data Voodoo Daddy Future M Preso Ed Alexander Oct 2012

391 views

Published on

A few chapters on some cool tools, analytic techniques, case examples and sales/marketing partnership opportunites.

  • Be the first to comment

Big Data Voodoo Daddy Future M Preso Ed Alexander Oct 2012

  1. 1. Agenda / Menu 1
  2. 2. Big Data Voodoo Daddy Agenda / Menu# BDVD # FutureM 2
  3. 3. Big Data Voodoo Daddy (or mama) Agenda / Menu # FutureM 3# BDVD 3
  4. 4. # BDVD Ed Alexaner Ed Alexander, Managing Consultant @fanfoundry Agenda / Menu# BDVD # FutureM 4
  5. 5. Agenda / Menu What is it? News and Views Cultural and consumer trends Corporate Trends Technology Landscape (the Cool Tool Pool) Demo Time A Test Methodology (BADIR) Use Cases Ways to test your own data Get Better Data (7 Quiz Questions) 5 Public Sector Mashups Get Real (time) Future Events, Resources # BDVD # FutureM 5
  6. 6. Defining “big data” – the four V’s: # BDVD # FutureM 6
  7. 7. If DataCouldTalk…# BDVD # FutureM 7
  8. 8. If ( ) DataCouldTalk…# BDVD # FutureM 8
  9. 9. Challenges – tooling up to: • Capture, combine and curate • Store, search and share • Analyze and visualize # FutureM 9 # BDVD 9
  10. 10. Opportunities • Internet search • Business informatics • Medical research • Genomics • Astronomy • Aviation • Meteorology • Finance # FutureM 10 # BDVD 10
  11. 11. Sources – 2 new quintillion bytes / day • Sensors • Mobile devices • Cameras • Microphones • Social graph – UGC # FutureM 11 # BDVD 11
  12. 12. The news, in general… The worst economic crash in 75 years A world economy with no place to hide “Always on” connectivity Widespread distrust of business Activist shareholders and special interest groups How does it impact your marketing agenda? # BDVD # FutureM 12
  13. 13. Big Data in the news… # BDVD # FutureM 13
  14. 14. Big Data in the news… Article upshot: Don’t blame Wal-mart The customer has all the power Example: Kroger (coupon response) • 70% of targeted • 3.4% of mass mailed Analysts & Techs Quoted: • Kantar Retail • Symphony IRI Group • Catalina Marketing Modiv Media’s “Scanit!” device • 89 Degrees # FutureM 14 # BDVD 14
  15. 15. What next? # BDVD # FutureM 15
  16. 16. What next? Special Big Data Issue # FutureM 16 # BDVD 16
  17. 17. The corporate view: big data in marketing Emerging stages – some business sectors have gone mainstream; Marketing is tooling catching up Mainly departmental - not much data integration or sharing Intuition based on business experience is still a driver; data analytics plays a supporting role Data challenges persist: accuracy, consistency, access, realtime Talent shortage - challenges business to apply results Culture’s role: orgs with a “culture of measurement “ succeed # BDVD # FutureM 17
  18. 18. The corporate view: big data in marketing Bloomberg Business Week Research Services # FutureM 18 # BDVD 18
  19. 19. The corporate view: big data in marketing Bloomberg Business Week Research Services # FutureM 19 # BDVD 19
  20. 20. The corporate view: big data in marketing1. CXOs now paying attention. Why?• Competition – lead, catch up, patch up PR• Predictive Intelligence – detect, adapt, seize opportunity• Optimization - don’t want to leave money on the table2. Elusive answers are suddenly more attainable• Operations, Sales, Marketing, Customer Care, R&D, etc.3. Transformation can now be justified with data• Train managers as analysts so they can produce and consume data• Rely on their business knowledge to interpret and act on data4. Priorities can be tuned• Identify top few “needle mover” opportunities and focus on them• Decision support can gain visibility based on proven results 20 # BDVD # FutureM 20
  21. 21. Cultural trend: Data-driven, custom communication # BDVD # FutureM 21
  22. 22. Cultural trend: Data-driven, custom communication1992: sad :(PointCastIntrusiveIn your faceOff-targetPoor quality # BDVD # FutureM 22
  23. 23. Cultural trend: Data-driven, custom communication1992: sad :( 2002: mad ):PointCast “Push sux”Intrusive SubversiveIn your face IntrusiveOff-target SpookyPoor quality Invasive # BDVD # FutureM 23
  24. 24. Cultural trend: Data-driven, custom communication1992: sad :( 2002: mad ): 2012: rad! :)PointCast “Push sux” I want my MDVIntrusive Subversive WelcomeIn your face Intrusive ExpectedOff-target Spooky PreferredPoor quality Invasive …but secured? *MDV: Massive Data Visualization # BDVD # FutureM 24
  25. 25. The new consumer demand: “I want my MDV”: We’re always on, and doing it now - • Showrooming • Facebooking • GPS navving • Socializing – Foursquare, Twitter, Instagram, etc. • Shopping & Banking • Customer care Cool tool • Audience & Community building • World blending (ex: QR, text, POS, Call CenterRetail, ecommerce, mobile # BDVD # FutureM 25
  26. 26. The new consumer demand: “I want my MDV”: Millenials are Digital Natives – mobile, social and always on They blur the lines between the digital and physical world They are less concerned about what’s going on with their data * By 2020, they will account for 50% + of retail spending Post-millenials are growing up digital * They seek trust, transparency and authenticity # FutureM 26 # BDVD 26
  27. 27. Corporate Trends # BDVD # FutureM 27
  28. 28. Big Datas Shifting Focus: Transaction > Engagement Personal Systems Analog Transaction Engagement Experiential Fulfillment Circa Pre-1950s 1950+ 2000+ 2005+ 2010+ Reliability & Continuous Sense and Agility and Intention Design Point stability improvement response flexibility driven Challenge Human Computing Social Contextual Individual Comm. Style Analog Systems Dictatorial Conversational Role tailored Personalized Multi-channel, Bionic, Social-led, UX Physical Machine based real time portable omni-media Time / space Speed Governed Just in time Real time Right time continuum Corporate & Personal, Reach Physical Corporate Value chains Internet one to oneInformation & structured Immersive Self-aware, Word of mouth Knowledge flows Knowledge records & data information embedded Social Tangentially Fundamentally Pervasively Ubiquitously Water cooler orientation social social social social Intelligence Human based Hard coded Business rules Predictive Pattern based Loyalty, Social Community & Examples assembly line Payroll, ERP, CRM reward, games, relationship social business context managementSource: R Wang & Insider Associates, LLC. # FutureM 28 # BDVD 28
  29. 29. # BDVD # FutureM 29
  30. 30. # FutureM 30# BDVD 30
  31. 31. Gartner: 72% have a “CMTO” today # BDVD # FutureM 31
  32. 32. http://www.emarketer.com/Article.aspx?R=1008909# BDVD # FutureM 32
  33. 33. ( What, no real time? ) 72% http://www.emarketer.com/Article.aspx?R=1008909# BDVD # FutureM 33
  34. 34. Technology Landscape (Cool Tool Pool) DAM SEO Email Testing & Search & PPC ads Marketing Optimization VIdeo Landing Site add-ins Web sites Pages Marketing E-commerce SM Ads Automation Webinars Targeting Display adsCRM Community Personalization SM marketing Call center B2B Data Multi-channel Gamification Analytics Mobile Databases Design Creative Chat Big Data Events Video ads Datasets PR APIs Surveys Collaboration Cloud Business Customer Loyalty Intelligence Experience Location Agile # BDVD # FutureM 34
  35. 35. Technology Landscape (Cool Tool Pool) # BDVD # FutureM 35
  36. 36. Stretch Goals for Cool Tools1. Rapid time to value - always on, omni-channel, user chummy for staff and customers2. Point and click customization - user-driven, brain dead simple3. 360 degree customer view – every salient data source linked, integrated and secure4. Real time visibility - instant refresh for all customer-facing and decision making (tactical) occasions5. Clean data - easy for all users to maintain, inspect and fix6. High adoption - self-training, guided navigation, less clutter7. Extended success – new & extended capability, new advantage8. Broad community - best / better practice sharing – each one teach one # FutureM 36 # BDVD 36
  37. 37. The payoff: central data + cool toolsStrategic Goals1. Boost productivity and efficiency• Centrally accessible, multichannel marketing data• Serves across addressable marketing channels• Easier to find and act on than data trapped in silos. 2. Reduce costs, improve marketing productivity Centralized multi-channel marketing data:• Improves ability to target and glean subscriber intelligence• Improves efficiency of data intelligence tasks• Improves organizational alignment 3. Enhance customer segmentation and personalization• Consistent view into multichannel customer data• Improve segmentation, 1:1 personalization, relevance # FutureM 37 # BDVD 37
  38. 38. The payoff: central data + cool toolsTactical goals• Campaign analytics and testing• Optimization, Acquisition, Lead Generation• Predictive Modeling – what is your killer niche?• Segmentation / Personae – who acts how?• Attribution precision – across channels, online and offline• Valuation of social media• Design testing (multivariate testing) • Websites • Emails • Offers • Messages # BDVD # FutureM 38
  39. 39. It’s Demo Time!# BDVD # FutureM 39
  40. 40. Framing the Discussion (Surprise!)It’s not about data & dashboards, it’s about culture & context.Ask: how can data help solve problems and guide decisions?1. Decide which challenges you’d like to address. Examples: reducing customer churn ● improving sales reducing inventory cost ● improving upsell / cross sell improving service ● improving user experience2. Develop a use case – customers, partners, departments, staff3. Run a pilot project – involve those end-users4. Invest in ways that will help meet your challenges. # BDVD # FutureM 40
  41. 41. A Test Methodology: BADIR Business Analysis Data Insights Recommend Question Plan Collection Solutions # BDVD # FutureM 41
  42. 42. A Test Methodology: BADIR Business Analysis Data Insights Recommend Question Plan Collection Solutions Sidebar: Use BADIR not only to test and report on data, but to vet those Cool Tools. Ask: Does that “cool tool” help break down silos? Does it support integration of processes and data? Okay, moving on… # FutureM 42 # BDVD 42
  43. 43. A Test Methodology: BADIR Business Analysis Data Insights Recommend Question Plan Collection SolutionsVague: Hypothesis: Specific: Choices: How do yourHow should I What business Only collect The right findings answerimprove my beliefs will we the data you methodologies the businessmarketing test, and how? need and techniques question?spend?Specific:How can Iidentifyunderservedcustomers? # BDVD # FutureM 43
  44. 44. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsVague: Hypothesis: Specific: Choices: How do yourHow should I What business Only collect The right findings answerimprove my beliefs will we the data you methodologies the businessmarketing test, and how? need and techniques question?spend?Specific:How can Iidentifyunderservedcustomers? # BDVD # FutureM 44
  45. 45. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsVague: Hypothesis: Specific: Choices: How do yourHow should I What business Only collect The right findings answerimprove my beliefs will we the data you methodologies the businessticket sales? test, and how? need and techniques question?Specific:How can Iidentifyproductiveticket salesinitiatives? # BDVD # FutureM 45
  46. 46. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsVague: Hypothesis: Specific: Choices: How do yourHow should I What business Only collect The right findings answerimprove my beliefs will we the data you methodologies the businessticket sales? test, and how? need and techniques question?Specific: Hypotheses:How can I 1. Will an early bird discount sell tickets?identify 2. Will a promo code help sell tickets?productive 3. Will a promo code stimulate referrals who buy?ticket sales 4. Will people still buy at full price?initiatives? Let’s analyze current data # BDVD # FutureM 46
  47. 47. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsVague: Hypothesis: Specific: Choices: How do yourHow should I What business Only collect The right findings answerimprove my beliefs will we the data you methodologies the businessticket sales? test, and how? need and techniques question?Specific: Hypotheses: QTY PCTHow can I 1. Will an early bird discount sell tickets? . . . . . . . . . 231 28%identify 2. Will a promo code help sell tickets? . . . . . . . . . . . 149 19%productive 3. Will a promo code stimulate referrals who buy? 262 32%ticket sales 4. Will people still buy at full price?. . . . . . . . . . . . . . 168 21%initiatives? 810 # BDVD # FutureM 47
  48. 48. Case Study #1: Data Insights Collection QTY PCT 231 28% 149 19% 262 32% 168 21% 810 # FutureM 48 # BDVD 48
  49. 49. Case Study #1: Data Insights Collection QTY PCT 231 28% 149 19% 262 32% 168 21% 810 # BDVD # FutureM 49
  50. 50. Case Study #1: Data Insights Collection QTY PCT 231 28% 149 19% 262 32% 168 21% Community 810 # FutureM 50 # BDVD 50
  51. 51. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsVague: How do yourHow should I findings answerimprove my the businessticket sales? question?Specific: QTY PCTHow can I 231 28%identify 149 19%productive 262 32%ticket sales 168 21%initiatives? Community 810 # FutureM 51 # BDVD 51
  52. 52. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsNext up:MultichannelattributionBehavioralScoring Hypotheses: QTY PCTSocial Sharing 1. Will an early bird discount sell tickets? . . . . . . . . . 231 28%impact 2. Will a promo code help sell tickets? . . . . . . . . . . . 149 19% 3. Will a promo code stimulate referrals who buy? 262 32%Geo/Pop/Wealth 4. Will people still buy at full price?. . . . . . . . . . . . . . 168 21% 810 # FutureM 52 # BDVD 52
  53. 53. Case Study #1: Business Analysis Data Insights Recommend Question Plan Collection SolutionsNext up:MultichannelattributionBehavioralScoringSocial SharingimpactGeo/Pop/Wealth # FutureM 53 # BDVD 53
  54. 54. Case Study #1: Business QuestionNext up:MultichannelattributionBehavioralScoringSocial SharingimpactGeo/Pop/Wealth # FutureM 54 # BDVD 54
  55. 55. Case Study #2: Catalog Retailers (national brands) # BDVD # FutureM 55
  56. 56. A Marketing Optimization Map PLANNING OPTIMIZATION WEB SERVICES ENGAGEMENT MOREM Analytics Optimization Response Internal CA Dashboards Management OR External NK S Reporting RequestE Chat U ManagementT M WebE E Offer ConsumerR Offer Portal Catalog Data R Messaging + Data Catalogs Adapters + Demos & Lifestyle + Life-Stage CUSTOMER ECOMMERCE + Purchase Behaviors DW SYSTEMS + Security & Preferences AND POS Enhancement Client Systems Data # BDVD # FutureM 56
  57. 57. Testing your data # BDVD # FutureM 57
  58. 58. Ways to test your own dataMultivariate Testing - testing more than one element of anoffer, website, email etc. in a live environment. Multiple A/Btests.Grail quest: optimize content across channels and contacts Content Contacts ChannelsLimits:• Time – to obtain statistically valid samples• Complexity – although tooling helps greatly• Computing power – although Cloud apps / hosting helps # BDVD # FutureM 58
  59. 59. Where to test?Online is easiest (but offline can be tested, too) Email: • Open, click & convert rates Website: • Landing page conversions • User registration pages • E-commerce checkout processes Offline: POS, Call Center, Catalog, Brochure, Signage, Layout # BDVD # FutureM 59
  60. 60. What to test?Effect or response to changes in Physical Appearance Elements• Copy• Layout• Images• Colors (backgrounds, etc.)Effect or response to changes in Content Elements• Price points• Purchase incentives• Premiums• Trial periods # BDVD # FutureM 60
  61. 61. Testing’s biggest challenge:Complexity – it happens quickly! Example: To test 3 different images in 3 different locations, you need to test how many possible combinations? a) 9 b) 18 c) 27 # BDVD # FutureM 61
  62. 62. Testing’s biggest challenge:Complexity – it happens quickly! Example: To test 3 different images in 3 different locations, you need to test how many possible combinations? a) 9 b) 18 c) 27 # BDVD # FutureM 62
  63. 63. Test toolsBrowser side (page tagging)Examples (visit www.whichmvt.com for more) :Server Side (DNS proxy, or hosted in your data center)Examples: # BDVD # FutureM 63
  64. 64. Test methodsDiscrete Choice / Choice Modeling (complex)Vary the attributes or content elementsQuantify impact of combinations on outcomesDiscover interaction effectsOptimal DesignIterations and waves of testingConsider relationships, interactions, constraints across elementsTaguchi MethodsReduce variations yet obtain statistically valid test results # BDVD # FutureM 64
  65. 65. Get better data # FutureM 65 # BDVD 65
  66. 66. 7 Quiz Questions for Better Data1. What data should I have? Look at your core mission, values, vision, strategy • What 5 things will impact the business in the coming year? o Ex: Will weather patterns affect L. L. Bean’s winter sales? • What are revenue drivers – quarterly, annually, channelwise? o Can new big data sources yield competitive advantage? • What are the “subjective” success criteria? Sales? CRV? Lift? Decide what matters, and set objectives from that. # BDVD # FutureM 66
  67. 67. 7 Quiz Questions for Better Data2. What metrics should I have? • Define Measurable goals - R&D, Marketing, Support, Sales, Ops, Finance, Engineering, HR etc. • Determine the right metrics. • Make certain you have the tools to measure them. # BDVD # FutureM 67
  68. 68. 7 Quiz Questions for Better Data3. What stands in the way? Get clarity and agreement on how to measure goal attainment. Example: “Better customer service” is a bit too nebulous • Metrics with inaccurate or incomplete data • Metrics that are complex or difficult to explain • Metrics that complicate operations or create excessive overhead • Metrics that cause people to act at cross purposes with the firm. An outsider should be able to audit if objectives were met. # BDVD # FutureM 68
  69. 69. 7 Quiz Questions for Better Data4. How can I get data and measurements on demand? SaaS apps can help you connect dataflow to analysis. Just beware the locked spreadsheet. • Salesforce.com: good for sales and dealflow • HubSpot: good for web marketing • Quickbooks, Excel: linked via xml app to data flow for instant financial / accounting updates and reports Departmental dashboards can enable weekly, daily, hourly or realtime trendspotting and fast course corrections. # BDVD # FutureM 69
  70. 70. 7 Quiz Questions for Better Data5. How can I empower everyone with on-demand insights? Create a Culture of measurement. • Maintain transparency to avoid surprises • Celebrate wins as they occur • Keep people properly motivated and on the same page Link rewards to the right performance measures All this makes it easier to work toward common, unified, clearly understood goals. # BDVD # FutureM 70
  71. 71. 7 Quiz Questions for Better Data6. Where to I start? Start at the top. • Set a strong example for people to follow • Publicize goals and keep your own progress visible • Demonstrate commitment to attaining shared goals • Pick the 5 most important goals and get the salient data Even if your targets were “off” at the outset, demonstrate success toward something, even if it’s just better intelligence. Pilot projects are learning labs. # BDVD # FutureM 71
  72. 72. 7 Quiz Questions for Better Data7. What should I do differently today? Continually question, re-evaluate and refine. • External factors can affect progress toward goals at any time. • External factors can affect goal setting at any time. • External factors can affect goal selection at any time. • Cultural factors can affect generation and use of data insights Determination is good, just keep it aimed productively. # BDVD # FutureM 72
  73. 73. Public SectorMashups # FutureM 73 # BDVD 73
  74. 74. 5 Public Sector Mashups1. Hurricane Risk Calculator Houston, TX Source: • NWS + historic data Use: • Neighborhood-level risk prediction http://risk.rtsnets.com • Predict flood, wind & power outages • Aids go/no go evacuation decisions # BDVD # FutureM 74
  75. 75. 5 Public Sector Mashups2. Quake-Catcher Network Stanford, CA Source: • Laptop accelerometer data http://qcn.stanford.edu Use: Improve on seismographic data • More location specific • Vastly cheaper • Free (laptop drop protection) • Easy to install in desktop PCs # BDVD # FutureM 75
  76. 76. 5 Public Sector Mashups3. Centers for Disease Control Atlanta, GA Source: • Google & Twitter search trends http://cdc.gov Use: • Speed disease detection • Enable response precision • Prevent & contain outbreaks • Eliminate SARS-like recurrence • Save lives • Support virality research # BDVD # FutureM 76
  77. 77. 5 Public Sector Mashups4. Predictive Policing Mountain View, CA Sources / mashup: • Foreclosures, school schedules, past crimes, bus schedules, library visits, weather conditions Use: • Predict likely crime occurrences • Focus police intervention efforts # BDVD # FutureM 77
  78. 78. 5 Public Sector Mashups5. Homeland Security Washington, DC F.A.S.T Module Sources: • Human suspect readings • Pulse, speech, CV, etc. • Bio, Interpol, other databases Use: • Predict malintent • Gather suspect intelligence # BDVD # FutureM 78
  79. 79. The world is your mashup Device / UI – web, mobile, social, print, POS, etc.Meta data – session info, device state, features, sensors Connectors, apps, processors, Cool Tools “plus” Mashup data – public, leased, licensedProprietary data – customers, partners, inventory, assets # FutureM 79 # BDVD 79
  80. 80. Get real (time) # FutureM 80 # BDVD 80
  81. 81. Real Time Direct Marketing Tools"Sales for Service" app Lead Nurturingcustomer interaction data from call ctr & POS Lead Scoringtailors offers quickly upon purchase / conversionimproves cross / upsell programs and offer targetingincludes: offer repository, biz rules engine, contacthistory DB, predictive analyticsTurns call center from a cost to a profit center (Email marketing) API to SFDC consolidates response in CRM(ID web visitors by IP)slices by: biz size, vertical, industry, geo(crowdsourced DBs) Find people and companiesTechprospex (ID tech used by B2B company) customer analyticsDrills down by model, version improves & automates sales response # FutureM 81 # BDVD 81
  82. 82. Real Time Direct Marketing Tools Persona triggers Lead Lists Marketing EmailAutomation Customer Analytics BI / Prospect Intelligence # BDVD # FutureM 82
  83. 83. Example: But now who owns it? Persona triggers Lead Lists Sales EmailMarketing Customer Analytics BI / Prospect Intelligence # FutureM 83 # BDVD 83
  84. 84. So, now who owns it? Marketing WWDDD ?Call centerCatalogEvent CommunitiesMobile ChannelsPOS CRM Support Storage,Print Integration,Social Service Access,Web Privacy, Sales Security IT # FutureM 84 # BDVD 84
  85. 85. Discuss, discuss Where is your data? Do you have a handle on it? Where does the data reside in your organization? Are there brilliant successes you can build on? Have you benchmarked your competitive space? Have you benchmarked a Disney-like experience? # FutureM 85 # BDVD 85
  86. 86. Future Events and ResourcesA DMA / NCDM Dec. 2012 Event # BDVD # FutureM 86
  87. 87. ReferencesTechAmerica FoundationPutting Big Data and Advanced Analytics to Work (McKinsey)The Logic behind Retailers’ Mercurial Pricing (HBR)The Current State of Business Analytics: Where do We Go from Here?(SAS / Bloomberg Business Week Research Services)Top 16 Tools to Create InfographicsTackling Multichannel Attribution (John Young, Epsilon)Predictive Analytics WorldTaming the Big Data Tidal Wave (Bill Franks, Teradata) # BDVD # FutureM 87
  88. 88. ResourcesAnalysis and Data Visualization Tools # FutureM 88 # BDVD 88
  89. 89. Thank you! .com +1 (781) 492-7638 USA East @fanfoundry 89

×